Software-Defined Automation Emerges as the Missing Link for Physical AI

KEYWORDS: Software-Defined Automation, Physical AI, Edge AI, Industrial AI, Machine Control, IT/OT Convergence, Digital Twin, Industrial Edge, Real-Time Control

Overview

Smaller, focused events can sometimes reveal emerging industry trends more clearly than large supplier-led conferences because they bring practical use cases, market observations, and technical debates into closer contact. Connection Day 2026, organized by Xentara, a start-up developing an open real-time edge platform for software-defined automation, did just that. Across sessions on industrial automation, AI, edge computing, and machine-level implementation, one theme stood out: industrial AI will scale only when intelligence can be connected to deterministic, real-time action on heterogeneous equipment.

Connection Day 2026, organized by Xentara, a start-up developing an open real-time edge platform for software-defined automation, revealed emerging industry trends more clearly than large supplier-led conferences.
 

Key Takeaways

• Software-defined automation is emerging as the practical link between industrial AI ambition and real-time machine execution.
• Physical AI, edge inference, and agentic workflows will require deterministic control, governed data models, and lifecycle-ready architectures.
• Early adoption will likely begin with targeted brownfield and machine-level use cases where flexibility, standardization, and reduced integration effort create measurable value.
• Manufacturers should treat software-defined automation as both a technology architecture and a transformation roadmap, with IT, OT, safety, and cybersecurity considered from the start.

Software-Defined Automation as the Bridge to Physical AI

Manufacturers face a widening gap between the intelligence now becoming available through AI and the installed automation base that must execute decisions safely, predictably, and economically. Traditional automation architectures were not designed for continuous model updates, semantic data models, local inference, agentic workflows, or rapid integration across machines and production lines. They were designed for deterministic control, long lifecycles, and vendor-specific engineering environments. The next phase of industrial transformation requires both forms of capability.

Several speakers framed this as a competitiveness issue. Patrick Ruthardt of Roland Berger argued that the next frontier of AI involves systems that perceive, decide, and act in the physical world. This requires more than conventional machine learning; it requires closed-loop architectures that connect data, training, deployment, inference, monitoring, and actuation. His message was not that AI replaces automation, but that Physical AI depends on software-defined manufacturing to make AI operational at scale.

Andreas Geiss of Xentara made the competitiveness argument more urgent by describing the convergence of AI/ML, IT/OT, and edge computing as forces that multiply rather than add. His core point was that German and European industrial depth remains a strategic asset, but it must be paired with speed. Software-defined architectures offer a way to preserve domain knowledge and engineering rigor while reducing the time required to adapt machines, integrate data, deploy intelligence, and scale solutions.

Anna Ahrens of Omdia compared software-defined automation with the software-defined vehicle journey, reinforcing the same point from a market-readiness perspective. Automotive has already shown that software-defined transformation is evolutionary, complex, and organizationally demanding. For manufacturing, the barriers are not only technical: she highlighted missing business models, infrastructure readiness, lifecycle liabilities, skills gaps, and the need to shift from capital expenditure thinking toward total cost of ownership. In other words, software-defined automation is not just a control technology; it is an operating model change.
 

ARC Advisory Group clients can view the complete report at the ARC Client Portal.

Contact Us if you would like to speak with the author.

Obtain more ARC In-depth Research Market Analysis.